When Does Self-supervision Improve Few-shot Learning?

2019-10-08 · ECCV 2020 8 · Jong-Chyi Su, Subhransu Maji, Bharath Hariharan

We investigate the role of self-supervised learning (SSL) in the context of few-shot learning. Although recent research has shown the benefits of SSL on large unlabeled datasets, its utility on small datasets is relatively unexplored. We find that SSL reduces the relative error rate of few-shot meta-learners by 4%-27%, even when the datasets are small and only utilizing images within the datasets. The improvements are greater when the training set is smaller or the task is more challenging. Although the benefits of SSL may increase with larger training sets, we observe that SSL can hurt the performance when the distributions of images used for meta-learning and SSL are different. We conduct a systematic study by varying the degree of domain shift and analyzing the performance of several meta-learners on a multitude of domains. Based on this analysis we present a technique that automatically selects images for SSL from a large, generic pool of unlabeled images for a given dataset that provides further improvements.

📄 PDF Abstract BibTeX arXiv:1910.03560

Code (2)

https://dagshub.com/arjun2000ashok/FSL-SSL

cvl-umass/fsl_ssl pytorch

Tasks

Few-Shot Learning Meta-Learning Self-Supervised Learning

Boosting Supervision with Self-Supervision for Few-shot Learning

2019-06-17 · Jong-Chyi Su, Subhransu Maji, Bharath Hariharan

We present a technique to improve the transferability of deep representations learned on small labeled datasets by introducing self-supervised tasks as auxiliary loss functions. While recent approaches for self-supervise…

Few-Shot Learning Self-Supervised Learning

Learning multiplane images from single views with self-supervision

2021-10-18 · Gustavo Sutter P. Carvalho, Diogo C. Luvizon, Antonio Joia Neto, Andre G. C. Pacheco 외

Generating static novel views from an already captured image is a hard task in computer vision and graphics, in particular when the single input image has dynamic parts such as persons or moving objects. In this paper, w…

SVL-Adapter: Self-Supervised Adapter for Vision-Language Pretrained Models

2022-10-07 · Omiros Pantazis, Gabriel Brostow, Kate Jones, Oisin Mac Aodha

Vision-language models such as CLIP are pretrained on large volumes of internet sourced image and text pairs, and have been shown to sometimes exhibit impressive zero- and low-shot image classification performance. Howev…

General Classification image-classification Image Classification Representation Learning

Boosting Few-Shot Visual Learning with Self-Supervision

2019-06-12 · ICCV 2019 10 · Spyros Gidaris, Andrei Bursuc, Nikos Komodakis, Patrick Pérez 외

Few-shot learning and self-supervised learning address different facets of the same problem: how to train a model with little or no labeled data. Few-shot learning aims for optimization methods and models that can learn …

Few-Shot Learning Self-Supervised Learning

Few-shot Semantic Segmentation with Self-supervision from Pseudo-classes

2021-10-22 · Yiwen Li, Gratianus Wesley Putra Data, Yunguan Fu, Yipeng Hu 외

Despite the success of deep learning methods for semantic segmentation, few-shot semantic segmentation remains a challenging task due to the limited training data and the generalisation requirement for unseen classes. Wh…

Few-Shot Semantic Segmentation Segmentation Semantic Segmentation